exa is a modern replacement for ls

A few suggestions to actually make this useful (in no particular order):
  • Rewrite in C
  • Don't advertise it as a 'replacement' for ls, this systemd attitude is scary
  • Have a command line switch which allows backward compatible / same output like ls
  • Choose another name (EXA is a graphics acceleration architecture of the X.Org Server)
  • Fix those issues (having a leap year bug is pretty dangerous, having it for 4 years already is ridiculously irresponsible)
 
Oh, we dont want "negativity" here? That reminds me... of certain very creepy esoteric self-awareness workshops... very spiritual and very enlightenend - and very dishonest, btw.

It's much more severe than just esoteric workshops: in our brave new world, disagreeing, or simply not applauding - i.e. being "negative" - has become a crime. It is true in the workplace, but is also more an more common in interpersonal, non-professional relationships.
 
  • Like
Reactions: a6h
I have no problem with exa, judging just by what little was shown here. I see it being in rust as no problem. If that makes the author(s) productive, good for them.

It doesn't replace ls. ls is still needed, partially for POSIX completeness, partially for use in scripts. Sure, one could make exa be a complete superset of ls ... but that would amount to just re-implementing the wheel, and this time getting it less round.

Tree viewers for directory structures are old and common. "Tree" has been around for ages. The fact that I haven't installed it in the last ~15 years tells me that I have no need for them.

I would not use exa, nor would I install it. But if others like it, I have no problem with that.
 
I'm sure many here use Go, and if it wasn't for Plan 9, alef, Inferno, and Limbo, there many never have been any Go.

I like Plan 9 but only the good parts.

Alef and Limbo lead the industry down a very dark path of an architecture necessitating huge bloated unportable VMs required to process bytecode. The (many the same) developers have realized that and are trying to undo the pain that they have caused by Golang going *back* to native and static compilation using (until semi-recently a modified ~C99 compiler (Ken's C Compiler from Plan 9)).

Without this error, we would not have had crudware like JVM, .NET or Flash.
So basically, sometimes it *is* better to do nothing at all.

In theory, without Plan 9, we may have had something like Go 10 years earlier because people would be focusing on language improvement rather than trying to maintain VMs or fix their shortcomings.

(Also the requirement of the mouse is another bit of evidence that Plan 9 didn't get everything right).
 
In theory, without Plan 9, we may have had something like Go 10 years earlier because people would be focusing on language improvement rather than trying to maintain VMs or fix their shortcomings.

Alef and Limbo were pet projects by rob, not crucial parts of the Plan 9 operating system development.

(Also the requirement of the mouse is another bit of evidence that Plan 9 didn't get everything right).

I disagree. The mouse is one of the best IT inventions from the 1980s and I find it disappointing that people still prefer to simulate teletype devices.
 
It is quite sad to see the negativity instead of encouragement
Encouragement should be for things that are dramatically different and useful. If one comes up with a mouse trap that snaps five percent faster, you won't get encouragement. Encouragement for the sake of encouragement is something you get from your Mom. Then saying it replaces the tried-and-true with minor improvements, then you're asking for negative comments.
 

As expected, some people can't help themselves. Say whatever you prefer because I won't read it.

Of course, as shown by the few responses already, there is no self reflection. Very pathetic.

Good luck. Please delete my account.
please...help me to help you..to help to others to think like you
nothing personal ,but if somebody insult me like you do (i'am a new member but feel part of the FreeBSD comunity)
you know the answers you will get
maybe if you write a tool that replace cp command because if cold and ugly for one that do:
-colored output (like the hackers movies)
-a web browser built in
-a file manager too
-a ssh server
-and make you cofee, but you need the plugin(is all writen in java,so it well be fast,why we need C ? is old and ugly)

but..please, follow the light and comeback to iluminate all
 
Alef and Limbo were pet projects by rob, not crucial parts of the Plan 9 operating system development.
That is good to hear. I always thought that Plan 9 had some great innovations but really dropped the ball when it came to these. Just 5c, 6c, 7c alone was a great idea (for handling more than intel architecture) when it came to language technology. Shame the industry didn't adopt the good part.

I disagree. The mouse is one of the best IT inventions from the 1980s and I find it disappointing that people still prefer to simulate teletype devices.

For consumer use, yes absolutely (including touchpad / tablet screen). I am also not using that term in an arrogant way. When I am off duty and just want to switch to a different film, I do *not* want to have to type anything to do so!

But for an environment where it is predominantly command line tools (which Plan 9 was by design), the mouse simply did not offer enough. It makes it hard to use in VMs, serial debugging. Even today it is awkward because you needed a true three button mouse (most laptops do not have that and requires emulation).

I am no usability expert and since we still have both approaches; there is probably no correct answer. But to enforce the use of a mouse (which Plan 9 did), is certainly not correct.
 
Shame the industry didn't adopt the good part.

The good part of Plan 9 was the perfection of its "everything is a file" scheme: The Acme "text editor" exposes a file system for its toolbar, another file system for its file buffer, yet another file system for the second buffer, ... - I wish 9P (or 9P2000, Styx or whatever) was more common today. I know that Microsoft has adopted some of it for its WSL2 system, there's also /proc, but that's about it, sadly.

One could debate about whether rio was a good part as well, though.

for an environment where it is predominantly command line tools (which Plan 9 was by design), the mouse simply did not offer enough.

Plan 9 follows a similar user interface as vi/Vim: The mouse is your "command mode", the keyboard is your "insert mode". I wonder why people understand Vim but fail to see the benefits of Plan 9.

Even today it is awkward because you needed a true three button mouse

Or one with a clickable wheel and a not-too-esoteric mouse driver. Which are most mice. :) Then again, if you don't have an actual mouse, you probably don't want to require one in the first place and Plan 9 might not be for you.
 
you probably don't want to require one in the first place and Plan 9 might not be for you.

Ugh, don't remind me. As part of my thesis I made two ports of software to Plan 9, a simple software renderer (I originally wrote on MS-DOS) and my network implementation of OpenGL (OpenGL|Distributed) so that I could measure some performance characteristics. It required a fair number of changes (especially in the networking system to use "dial" and other Plan 9 facilities) so I had to pick up acme which I found fairly nice (it was cool how text could be piped to commands) even if it looked a bit messy with my "saved commands" in the toolbars etc.

I really wanted to like Plan 9 and for the underlying system, I really found the architecture elegant, it was really clean. I even liked the UI and how minimal it was. However, what it did not have was a proper terminal and that crucified my productivity. No history, etc. (I couldn't even use arrow keys in acme to go up a line, even though that was not terminal related).

In the end I had to compile up the Vim port and use that for pretty much everything; it provided a better terminal emulation than Plan 9. I did manage to avoid APE however and all the code was the Plan 9 way. Including <u.h> at the top of most compilation units was weird though haha.

Screenshot of possibly one of the few 3D programs you will see running on Plan 9. Compared to Doom (or Quake) this did not run fast however (in pure software mode) ;).
 

Attachments

  • qsoft_plan9.png
    qsoft_plan9.png
    47.3 KB · Views: 209
Cheers, I will look into that. It is a slight shame it drags in plan9port which is a little large just for an editor but on many of my development machines I have that installed anyway ;)

A lighter alternative is Wily (https://en.wikipedia.org/wiki/Wily) but unfortunately is not quite as featureful as acme and had an X11 bug that I couldn't quite manage to fix.
 
In my opinion, Wily makes no sense anymore when you already have plan9port. If you prefer a low overhead, Acme2k might be an alternative as a drop-in replacement for Acme. (But, unlike Acme, it still doesn’t support files with spaces in their paths.)
 
In my opinion, Wily makes no sense anymore when you already have plan9port. If you prefer a low overhead, Acme2k might be an alternative as a drop-in replacement for Acme.

Yes, that is very true.
Acme2k now supports cursor movement using keyboard (my only gripe with standard Acme), so I will certainly try this and edwood out.
 
I think the majority of the negativity comes from the way this was presented.. "A modern replacement"? That also makes me go "whatever..." because I can't help pick up some arrogance in there as well. Probably not intended and I'm well aware that I might be oversensitive there but yah... "It's a modern alternative for ls" would have made this sound a lot milder.

Sure, this is nitpicking. But what would you expect after seeing some dweebs (= personal opinion) trying to replace the foundation of Linux / Unix through systemd?

Just my 2 cents here of course.
 
Back
Top